1. Field of the Invention
The present disclosure relates to an image-reading apparatus.
2. Description of the Related Art
There has been a desire to convert some media, such as a single photograph including a film photograph generated by photographing using a film camera, or a printed photograph obtained by printing based on a piece of image data, or a photograph album (album) including a plurality of photographs, into image data and store the converted image data in a hard disk or other recording media, or in a cloud computing environment. Image data is typically generated by reading media with an image-reading apparatus (e.g., a flatbed scanner, an automatic document feeder (ADF) scanner, an overhead scanner, a hand-held scanner, and a digital camera).
With a flatbed scanner, a user places one or more sheets of a single photograph, or an area of a photograph album including at least a photograph that one wishes to scan on a glass plate of the scanner. An image capturing unit of the scanner captures an image of a reading surface on the glass plate through the glass plate. With an ADF scanner, the image capturing unit can consecutively capture an image of a plurality of photographs. With an overhead scanner (refer to Japanese Patent Application Laid-open No. 2001-268323) or a digital camera, a user places one or more sheets of a single photograph, or an area of a photograph album including at least a photograph that one wishes to scan within an image-capturing area of an image capturing unit. The image capturing unit captures an image of a reading surface facing the image capturing unit. With a hand-held scanner (refer to the publication of US Patent Application No. 2011/0142371), a user holds the hand-held scanner and moves the scanner while the scanner is in contact with the single photograph or the photograph album to enable an image capturing unit of the scanner to capture an image of the whole area of the photograph.
With an image-reading apparatus that reads a medium one by one such as an ADF scanner, an image of the medium is read by an image capturing unit that captures the image of the medium on a back pad in a monotone color. The back pad in a monotone color is used to obtain an image of the medium having a background in a uniform monotone color such as white or black. Subsequently, image processing is performed on an image of the medium being read, thereby generating final image data. In the image processing, an edge detection is performed in which the borderline between the medium and the background is detected in the read image of the medium from both top to bottom, and left to right. After that, the linearity of the detected edge is obtained, thereby locating the medium. Then, the image including only the medium without the background is cropped (refer to Japanese Patent Application Laid-open No. 2007-88654).
As for an edge detection of image processing, a technique has been reported, in which an image-reading apparatus reads an image of a medium using a line sensor such as a charge coupled device (CCD). The image-reading apparatus detects a shadowed area generated due to irradiation of the medium having an unevenness in a thickness direction with light emitted from a light source, through detecting the asymmetry of the inclination of an edge. When the image-reading apparatus determines the shadowed area, image processing is performed to eliminate the determined shadowed area, or the shadowed area and some pixels at or near the shadowed area (Japanese Patent Application Laid-open No. 10-285377).
It is also desired to prevent the medium from being damaged, readily convert the medium into image data, thereby generating high-quality image data. Actually, with a flatbed scanner, an ADF scanner, or an overhead scanner, a user has to bring the medium to the position where the scanner is installed and dispose the medium so as to be captured by the image capturing unit. With an ADF scanner, for example, before converting a photograph from a photograph album into image data, the user first has to take out the photograph from the photograph album. When using a photograph album in which a photograph is fixed with an adhesive, the photograph is damaged when being separated from the album. In addition, once a photograph is separated from the album, the adhesive force of the adhesive decreases, resulting in the difficulty of re-attaching the photograph to the photograph album.
With an ADF scanner, a photograph is conveyed by a carriage roller to the image capturing unit, resulting in damaging the photograph. In the same manner, with a hand-held scanner, the medium comes into contact with the scanner, and the medium is likely to be damaged.
In addition, when converting a single photograph into image data using a hand-held scanner, the user has to hold the photograph by hand, for example, to fix the photograph. Furthermore, the user has to move the hand-held scanner, resulting in the user to perform more work, which bothers the user. Since the user moves the hand-held scanner on the medium, a relative velocity between the medium and the hand-held scanner becomes variable, which makes it difficult to maintain the quality of the generated image data at a high level.
When using an overhead scanner or a digital camera, the photograph or the photograph album is exposed to the outside during image-capturing, and thus effected by the external environment. A photographic paper used for film photograph or a glossy paper used for a printed photograph has a higher reflection factor in comparison with plain paper. Ambient light emitted from the outside to the medium is therefore reflected on the reading surface of the medium and then is incident on the image capturing unit. This results in an inconvenience caused by an increase of the reflected ambient light which may be incident on the image capturing unit, i.e. reflections of the ambient light are generated, which makes it difficult to maintain an image level of the image data generated at high level.
In a conventional image-reading apparatus, although, cropping an image of the medium being read is performed by focusing on an edge component or a color component of an image, it is difficult to crop an image of the medium with high-accuracy according to this method. Therefore, further improvement is required.